Cette thèse porte sur l'étude de modélisation stochastique de carnet d'ordres, et de deux problèmes de contrôle stochastique dans un contexte de risque de liquidité et d'impact sur le prix des actifs. La thèse est constituée de deux parties distinctes.Dans la première partie, nous traitons, sous différents aspects, un modèle markovien de carnet d'ordres.En particulier, dans le chapitre 2, nous introduisons un modèle de représentation par profondeur cumulée. Nous considérons différents types d'arrivées d'évènements avec une dépendance de l'état courant.Le chapitre 3 traite le problème de stabilité du modèle à travers une approche semi-martingale pour la classification d'une chaîne de Markov dénombrable. Nous donnons, pour chaque problème de classification, une calibration du modèle à partir des faits empiriques comme le profil moyen de la densité du carnet d'ordres.Le chapitre 4 est consacré à l'estimation et à la calibration de notre modèle à partir des flux de données du marché. Ainsi, nous comparons notre modèle et les données au moyen des faits stylisés et des faits empiriques. Nous donnons une calibration concrète aux différents problèmes de classification.Puis, dans le chapitre 5, nous traitons le problème de liquidation optimale dans le cadre du modèle de représentation par profondeur cumulée.Dans la deuxième partie, nous proposons une modélisation d'un problème de liquidation optimale d'un investisseur avec une résilience stochastique. Nous nous ramenons à un problème de contrôle stochastique singulier. Nous montrons que la fonction valeur associée est l'unique solution de viscosité d'une équation d'Hamilton-Jacobi-Bellman. De plus, nous utilisons une méthode numérique itérative pour calculer la stratégie optimale. La convergence de ce schéma numérique est obtenue via des critères de monotonicité, de stabilité et de consistance. / This thesis deals with the study of stochastic modeling of limit order book and two stochastic control problems under liquidity risk and price impact. The thesis is made of two distinct parts.In the first part, we investigate markovian limit order book model under different aspects. In particular, in Chapter 2, we introduce a model of cumulative depth representation. We consider different arrival events with dependencies on current state. Chapter 3 handles the model stability problem through a semi-martingale approach for the denumerable Markov Chain classification. We give for each problem a model calibration from empirical facts such as mean average profile of limit order book density. Chapter 4 is dedicated to the model estimation and model calibration by means of market data flow. Thus, we compare our model to market data through stylized facts and empirical facts. We give a concrete calibration to the different stability problems. Finally, in Chapter 5, we handle the optimal liquidation problem in the cumulative depth representation model framework.We study, in the second part, an optimal liquidation problem of an investor under stochastic resilience. This problem may be formulated as a stochastic singular control problem. We show that the associated value function is the unique viscosity solution of an Hamilton-Jacobi-Bellman equation. We suggest an iterative numerical method to compute the optimal strategy. The numerical scheme convergence is obtained through the monotonicity, stability and consistancy creteria.

Cette thèse s’interroge sur les notions d’aléa et de régularité des variations boursières. Nous démontrons sur le plan théorique, la compatibilité des principales théories financières (cf. efficience informationnelle, finance comportementale et approche conventionnaliste) avec l'impossibilité de battre la stratégie «buy and hold». Cette impossibilité est confirmée par les études statistiques dans la mesure où les régularités identifiées dans les séries financières ne permettent pas de prédire le sens des variations futures. Les modèles économétriques disponibles à présent offrent souvent un «hit score» insuffisant (<60%) pour réussir des tentatives fructueuses de «market timing». Une contribution de ce travail se trouve dans l'introduction du concept de complexité algorithmique en finance. Une approche générale est proposée pour estimer la «complexité de Kolmogorov» des séries de rentabilités: après un processus «discrétisation-effacement», des algorithmes de compression sans perte sont utilisés pour détecter des structures régulières qui ne sont pas toujours visibles aux yeux des tests statistiques. En étudiant le degré d'aléa des principaux marchés internationaux à une fréquence «tick-by-tick», on constate une complexité plus élevée pour Euronext-Paris que pour le NYSE et le NASDAQ. Nous expliquons ce résultat par une auto-corrélation plus élevée des volatilités inter-journalières aux Etats-Unis qu'en France. L'inefficacité de «market timing» étant soutenue aussi bien par les théories financières que par les observations empiriques, nous définissons la notion de «battre le marché» dans ce sens spécifique avec un modèle mathématique qui s'inscrit dans le cadre de la calculabilité. / This doctoral dissertation examines different notions of financial randomness and regularity. We show that main financial theories (i.e. market efficiency, behavioral finance and the so-called ``conventionalist approach'') support the impossibility of outperforming the ``buy and hold'' strategy. This point is confirmed by statistical works since regularities identified in financial time series do not help to predict the direction of future returns. To the best of our knowledge, available econometric models often provide too low ``hit scores'' (< 60%) to become successful trading rules. A conceptuel contribution of this work lies in the introduction of algorithmic complexity to finance. A general approach is proposed to estimate the ``Kolmogorov complexity'' of financial returns: lossless compression tools are used to detect regular patterns which could be overlooked by statistical tests. By studying tick-by-tick data from major stock markets, we find a higher complexity for the Euronext-Paris data than for the NYSE and the NASDAQ ones. This result can be explained by their intraday volatility autocorrelations. Supported both by financial theories and by empirical observations, impossibility to outperform the ``buy and hold'' strategy is linked to the common expression ``to outperform the market'' by a new definition for ``unbeatable strings''. With computable functions modeling effective trading rules, a price sequence is said to be ``unbeatable'' if no effective trading rule can generate indefinitely more profits than the ``buy and hold'' alternative.

La présente étude consiste en l’analyse comparative de l’Autorité des Marchés Financiers et du Comité d’Etat de la Bourse, les autorités de régulation actuelles des marchés de capitaux de France et du Vietnam, deux pays qui partagent un héritage historique commun et des relations particulières dans de nombreux domaines, dont celui juridique. Non seulement le CEB et l’AMF connaissent des mutations dues au contexte économique global, mais le régulateur vietnamien a vu sa charge de travail s’accroître considérablement suite à l’accession du Vietnam à l’OMC. Organes d’encadrement voulus par les pouvoirs publics de la France et du Vietnam, le Comité et l’Autorité assurent la protection de l’épargne et veille à l’équilibre des marchés, en exerçant leurs attributions en matière de contrôle et en recourant à leur pouvoir normatif et d’élaboration d’actes non-décisoires de manière constante. Si leurs relations avec les autres entités du système financier varient sensiblement pour chacun d’eux, les disparités de leurs pouvoirs répressifs respectifs s’estompent de plus en plus pour aller dans le même sens. / This study consists of the comparative analysis between the Autorité des Marchés Financiers and the State Securities Commission, the current regulatory authorities of the capital markets in France and Vietnam, two countries which have a common historical heritage and particular relations in many fields, even legal industry. Not only the SSC and the AMF know changes due to global economic context, but also the Vietnamese regulator had its workload considerably increased following accession for Vietnam. Being management bodies which creation was wanted by public authorities in France and in Vietnam, the Commission and the French authority ensure the saving protection and attend to the market equilibrium, by performing their supervisory powers and by using their normative capacities and developing non-constraining acts in a constant way. If their relations with others entities of the financial system vary appreciably for each one, the disparities of their respective repressive capacities decrease and go more and more in the same direction.

This dissertation argues on the necessity for an efficient and credible recovery and resolution framework for central clearing counterparties (CCPs).Following the 2008 subprime crisis and the failure of Lehman Brothers, several reforms have been undertaken in order to reinforce the strength of the financial system and, in particular, the over the counter (OTC) derivatives market. Two streams of reforms are of particular relevance in the present context: a) The development of resolution regimes for financial institutions, in particular for systemic “too big to fail” banks. The key objectives of these resolution regimes are to ensure the continuity of the critical functions performed by systemic institutions, the preservation of financial stability and the protection of taxpayers (i.e. avoiding public bailouts). This stream completed regulatory initiatives to reinforce the existing prudential framework acknowledging the natural limits of the latter and the observed forbearance by supervisory authorities in the aftermath of the financial crisis. The academic literature analysing the development and implementation of resolution regimes for banks essentially looked at issues such as the respect of the shareholders’ fundamental rights, the existence of appropriate safeguards for creditors and the execution of bail-in operations. b) The obligation to clear centrally OTC derivatives markets in order to prevent and control the possible contagion of risks across the financial system. The objectives of this stream of reform are to increase the transparency of the OTC derivatives markets and reduce the counterparty credit risk in the system by forcing transactions through professional risk managers, i.e. central clearing counterparties (CCPs). This however assumes that CCPs have the capability and the capacity to manage this centralisation of risks. Therefore, the literature showed interest in the risk management practices of CCPs, in particular margining methodologies, and agency problems created by the ownership structure and the loss allocation mechanisms of CCPs, in particular analysing the effects of a CCP’s skin-in-the-game (i.e. own contributions to the loss allocation mechanisms) on the incentives of shareholders and clearing members. Although recent history has been rather merciful in terms of CCPs' failures, the inherent cross-border and international nature of their activities, their level of interconnectedness with other financial institutions and their growing importance following the G20 commitments on OTC derivatives make them critical nodes of the financial system. It is therefore important to consider carefully the recovery and resolution of CCPs that actually stands at the crossroad of the two streams of regulatory reforms mentioned above and completes them. Similarly, this dissertation intends to build the bridge between the literature on CCPs and the literature on banks’ recovery and resolution. It aims at providing a better understanding of the resolution framework developed for CCPs, associated policy choices and possible issues with a specific insight on the situation in Europe. / Doctorat en Sciences économiques et de gestion / info:eu-repo/semantics/nonPublished

L'introduction récente, dans la littérature financière, d'agents aux comportements imparfaits permet une représentation des marchés d'actions plus proche de la réalité. Nous pensons surtout au concept de bruiteurs qui insiste sur l'impuissance de certains agents rationnels à posséder et utiliser toute l'information disponible et pertinente. Cette thèse consiste en une description de l'impact de ces agents sur les marchés financiers et la recherche d'outils adéquats pour mesurer les conséquences de leur présence sur l'efficience des marchés. La première vision théorique des marchés financiers composés de froids calculateurs parfaitement rationnels et homogènes laisse la place à l'intégration d'agents aux anticipations hétérogènes. Les changements intervenant dans les hypothèses de comportements se sont accompagnés d'une modification du modèle testable dérivé (marche aléatoire, martingale). Mais tout cela s'effectue au sein de ce que nous appelons le paradigme probabiliste de l'investisseur rationnel même si l'introduction de modèles non linéaires (ARCH ...) permet une avancée significative. Le chaos déterministe offre une autre voie de développement de la non linéarité qui permet en plus de se libérer de la contrainte inhérente à l'emploi de modèles probabilistes. Nous montrons qu'une révision des tests d'efficience découle de ces techniques du chaos avec notamment l'emploi du test BDS que nous interprétons ici comme un test de l'efficience “classique” des marchés contre une hypothèse que nous appelons Efficience Bruitée. L'Efficience Bruitée définit un marché où coexistent agents rationnels parfaitement informés et bruiteurs. L'étude empirique menée sur le marché français des actions semble confirmer le bien fondé de notre concept. Elle met à jour les insuffisances des tests et modèles qui reposent sur la linéarité pour décrire la grande diversité des investisseurs présents sur les marchés financiers.

The three chapters all address solvency regulation issues, with a focus on market risks under the Solvency II framework. Chapter 1 deals with “high-level” aspects of Solvency II as main principles and the general structure. Chapters 2 and 3 will be devoted to quantitative issues. Chapter 1 describes the main evolutions that led to the development of Solvency II. The insurance sector has dramatically evolved during the last two decades. Among others developments, we stress the new risks faced by the sector as natural catastrophes, changing demographics or market risks. Insurers become international companies, investing almost 10 trillion € of assets in Europe at the end of 2014 and being increasingly intertwined with banks and other financial sectors. Financial innovation and the refinement of risk management techniques and models developed by companies have gained momentum among the major European insurance companies. Have these evolutions changed the needs for the supervisory of insurance companies? The economic foundation for regulation is based on the presence of market failures, including severe asymmetric information problems and principal-agent conflicts. Insurance consumers, particularly individuals and households, face significant challenges in judging the financial risk of insurers. But the importance of the insurance sector for financial stability has been increasing. A sound regulatory and supervisory system is necessary to maintain efficient, safe, fair and stable financial markets and promote growth and competition in the insurance sector. The difficult conditions experienced by the industry and the shortcomings of the previous regulatory and supervisory framework have forced regulators to take action to change the way in which they regulate insurance companies’ solvency. Recognizing the shortcomings of Solvency I, EU policy-makers undertook the Solvency II project. Solvency I was not consistently applied throughout EU as the directive allowed countries to implement insurance regulation in different ways. Moreover Solvency I did not consider risks fully or in detail. In life business, the major criticism was the lack of consideration of asset risks. Allowances for latest developments in risk management were also inadequate and companies could not use an internal model to calculate the solvency capital. Finally, the increasing presence of conglomerates and groups forced the insurance regulator to align some requirements with the banking regulation, Basel II/III. Due to the differences in their core business activities, banks and insurers regulators’ goal does not imply comparability of the overall capital charges. However, considering the asset side of the balance sheets, the investment portfolios of banks and insurers contain the same asset classes. In order to avoid regulatory arbitrage, the capital charges for the same amount and type of asset risk should be similar. Chapter 2 compares the main regulatory frameworks in Europe: Solvency II and the Swiss Solvency Test, SST, in Switzerland, with a focus on potential market implications. Both systems are quite advanced but some key differences need to be highlighted, including the treatment of assets, in particular sovereign bonds, the consideration of diversification or the risk measure applied. Solvency II uses a Value at Risk at 99.5% whereas the SST is based on a Tail Value at Risk at 99%. Our approach is both qualitative and quantitative. In particular, based on a numerical example, we aim at quantifying the level of regulatory capital prescribed by the standard models. The numerical analysis reveals large differences between capital charges assigned to the same asset class under Solvency II and the SST. Solvency II penalizes investment in stocks, mainly due to a lower diversification benefit under the standard formula. On the other hand the SST model requires a higher capital for bonds, primary due to a stringent risk measure and confidence level. The treatment of EU sovereign bonds under Solvency II is another area of concern as it does not require any capital for spread risk. The question arises to what extent an internal model leads to different capital requirements as compared to the SST and Solvency II models. Therefore we apply an internal approach based on Monte Carlo simulation to derive the necessary capital based on the Value at Risk at 99.5% (in line with the Solvency II standard model) and on the Tail Value at Risk at 99% (in line with the SST standard model). Internal models calculate capital requirements that more closely matches risks of insurers and promote a culture of risk management. To develop internal models, companies need incentives to properly manage their risks, i.e. decreasing capital requirements. One potential benefit of the standard model is that insurers who use it can be compared to one another, whereas internal models are by definition specific to individual insurers. One argument against the standard model is the possibility of some systemic risk. An unusual event in the capital or insurance market could encourage all insurers to take the exact same response, thereby causing a run in the market. The analysis shows that standard and internal models still display large discrepancies in their results, suggesting a long way ahead to achieve a harmonized view between the regulators and the insurance sector. The choice of a statistical model or the refinement of parameters are key concepts when setting up an internal model and appear to be critical in the Solvency Capital Requirement calculation. By calculating and comparing the market risk capital charges for a representative insurer under the Solvency II and the SST standard approach as well as an internal model, we are able to provide evidence that the regulatory framework might have an impact on asset portfolios. The main impacts would be a shift from long-term to shorter-term debt, an increase in the attractiveness of higher-rated corporate debt and government bonds, in particular EU sovereign bonds as the consequence of the special treatment under Solvency II, as well as low level of equity holdings. But it is unlikely that large-scale reallocations will happen in the short term, as transitional arrangements are likely to phase in the implementation of Solvency II over several years. The likely impact on assets portfolios could have also already been anticipating by insurers. Chapter 3 studies the effectiveness of the Solvency II reform to prevent the default probability faced by a life insurance company. The default risk leads to a consequence that policyholders might not get back their initial investment upon default of the insurance company. Therefore, policyholders are concerned with the issues like what probability the insurance company will become bankrupt and which amount they can expect to obtain after taking account of the default risk of the insurer. Starting from a theoretical life insurance company which sells a participation insurance policy containing only a savings component and a single premium inflow, we simulate a life insurance company on an eight-year time horizon. We focus only on market risks as there is no mortality risk attached to the insurance contract. Finally several policies and investment strategies will be analysed. The purpose of the chapter is to evaluate how Solvency II can prevent the company to collapse. The papers discussing Solvency II effectiveness are qualitative in nature. In particular there is little research on the accuracy of the standard formula with regard to the proclaimed ruin probability of 0.5% per year. To do so we compare the probability of default at maturity of the life insurance policy, i.e. if the company has to enough assets to pay what was promised to the policyholders, with the early probability of default forced by Solvency II based on standard and internal models. We have first to calculate the Solvency Capital Requirement as laid down in the directive. One crucial point is the evaluation of liabilities. To do so we use an approach recently applied by the insurance sector called Least-squares Monte Carlo (LSMC). The aim of Solvency II is to monitor insurers on an annual basis. The SCR level can then be interpreted as a regulatory barrier, consistent with a model developed by Grosen and Jørgensen (2002). Key drivers of the ruin probability at maturity include interest rate parameters, portfolio riskiness and investment strategies in bonds. The continuously decrease of interest rates creates a challenge for insurers, especially life insurers that suffer a double impact on their balance sheet: a valuation effect and a decreasing reinvestment returns of premiums and maturing bonds. The latter explain also the riskiness of rolling-bond strategies compared to duration matching strategies. By setting the confidence level to 99.5% per year, the regulator wants to ensure that the annual ruin probability equals to 0.5%. Since the SCR from our internal model equals the 0.5% quantile of the distribution, it exactly matches the targeted ruin probability. Our analysis reveals that the set-up and calibration of the Solvency II standard model are inadequate as the solvency capital derived by the standard formula overestimates the results of the internal model. This is mainly the consequence of an overestimated equity capital and a lower diversification benefit. The 0.5% proclaimed goal under Solvency II is not reached, being too conservative. One declared goal of the directive is to decrease the duration gap between assets and liabilities. Solvency II penalizes then rolling-bond strategies. The long-term feature of our policy should impact the level of regulatory capital. As Solvency II is based on a quantile measurement, we define the solvency capital using the default probability objective for different horizons. SCR is not systematically a decreasing function of the time horizon even if a decreasing form appears on long-term. This shows undoubtedly that a horizon effect exists in terms of measurement of solvency. As the standard model overestimated the internal model capital we expect a forced default probability higher than 0.5% under the Solvency II framework. The SCR barrier stops the company more often than it should be. This can be interpreted as one cost of regulation, i.e. closing down financially sound at maturity companies. The analysis of the evolution of default probabilities as a function of time horizon reveals that ruin probabilities at maturity lie always below the Solvency II objective. Furthermore the gap between the observed default at maturity and the Solvency II objective is increasing over time; the situation is even worse for longer-term insurance products. Finally stakeholders are more interested in their expected return than in the default probability. A cost of regulation defined as the difference between stakeholder’s returns with and without regulatory framework exists, particularly for shareholders. / Doctorat en Sciences économiques et de gestion / info:eu-repo/semantics/nonPublished

La persistance des fraudes financières depuis les origines de la création de la bourse conduit au constat selon lequel la sanction est un élément fondamental du bon fonctionnement des marchés financiers. En l’état actuel du droit positif, notre système de sanction repose pour l’essentiel sur l’Autorité des marchés financiers. À première vue, le droit français offre le visage d’une architecture moderne. Pourtant, l’actualité n’en finit pas de démontrer l’existence d’affaires et de fraudes sensibles affectant drastiquement la confiance des investisseurs dans le système. En vue d’une amélioration de notre système de sanction, la comparaison avec le droit canadien se révèle très enrichissante. Il en ressort que le système de sanction en droit des marchés financiers français pourrait être rendu plus efficace par la création d’une juridiction spécialisée à laquelle serait confiée l’ensemble du contentieux relatif aux marchés financiers. Une telle réforme favoriserait l’harmonisation et la cohérence du système de sanction. / The persistence of financial frauds since the creation of the stock exchange has made it clear that sanctions are a fundamental aspect to ensure the effective functioning of financial markets. Currently, our sanction system relies, for the most part, on the action of the financial regulator, the Autorité des marchés financiers. At first sight, French law offers a very modern system. However, frauds keep on happening, affecting drastically the trust of the investors in the system. In order to improve our sanction system, the comparison with Canadian law has revealed to be very rewarding. It is possible to make our system more efficient by implementing a special jurisdiction, which would be responsible for all the financial markets disputes, in order to make our sanction system more coherent and unified.

Thèse numérisée par la Direction des bibliothèques de l'Université de Montréal.

Comportement de l'investisseur et impact sur les prix du marché. / Sir Isaac Newton, who is one of the most influential physicist and mathematician of all time, after he suffered huge losses in tulip market said: “I can calculate the motions of heavenly bodies, but not the madness of people.” Financial markets are full of uncertainties. The movement and volatility in stock prices has been the focus of attention for scholars all the time. Over the last decades, financial markets gain influence both at people’s life and country’s economics as a result of technological advances, financial liberalization, and ongoing international trade. On one hand, participant’s property and investor’s market performance are impacted by price fluctuation. On the other hand, the development of national economic is closely interrelated to the stability of financial markets. In this effect, the understanding of investors’ designing making and how it affect the market price movement is of vital interest to both researchers and economic policy market. Experimental Finance has already become a well-established field, a fact that was recognized by the attribution of the Nobel Prize in Economics to Vernon Smith in 2002 who’s most significant work was concerned with market mechanisms and tests of different auction forms. However so far the major part of experimental work in Finance has considered (including Vernon Smith) human rationality and the ability of markets to find the proper price close to an equilibrium setting. [...]

Si la consommation, dans son acception économique, constitue le dernier stade du processus de production et de distribution des produits, elle est perçue, dans une acception juridique, comme étant le point de départ d‟un autre processus qui lie le consommateur au professionnel. Cette phase constitue le point de départ d‟un rapport qui fait naître des obligations au professionnel qui sont autant de droits pour le consommateur.Depuis le début des années 90, une attention particulière est consacrée à l‟épargnant et à sa protection, à travers d‟une part, des dispositions propres au droit financier, notamment par la régulation des marchés financiers, le contrôle des prestataires de services d‟investissement, et d‟autre part, des dispositions consacrées par le droit de la consommation notamment l‟encadrement du démarchage, de la commercialisation des produits et des services financiers, le renforcement de l‟obligation d‟information due par l‟intermédiaire financier et l‟extension de son obligation de conseil.L‟analyse juridique du processus de consommation des produits financiers passe par la combinaison de tous ces éléments. D‟une part, avant et lors de la formation des liens, les pouvoirs publics tentent de rétablir un équilibre dans les rapports entre professionnels et consommateurs. Ces derniers, étant réputés faibles économiquement, ne disposent pas des mêmes informations sur les produits et services financiers, sont sollicités par différents moyens et techniques dont ils n‟ont pas la maîtrise, et doivent faire des choix entre des produits variés, complexes et plus ou moins risqués. D‟autre part, lors de l‟exécution des prestations, à côté des risques liés aux produits, ou de l‟inexécution des obligations contractuelles, d‟autres risques peuvent naître des défaillances dans le fonctionnement du marché et des comportements de ses acteurs. Par conséquent, un renforcement des moyens de protection des consommateurs s‟avère nécessaire pour pallier aux insuffisances des moyens actuels dont ces derniers disposent pour obtenir réparation du préjudice et demander indemnisation. / If consumption, in the economic sense, is the final stage of the production and distribution of goods, it is perceived in a legal sense, as the starting point of another process linking the consumer to professional. This phase is the starting point of a report which gives rise to obligations which are all professional fees for the consumer.Since the beginning of the particular 90s, an attention is devoted to protection, through some measures specific to the financial law, including the regulation of financial markets, supervision of providers investment services, and other provisions that tend toward a reconciliation of the provisions enshrined in the consumer law including the regulation of canvassing, measures to provide protection in the marketing of products and services Financial distance, enhanced disclosure requirements owed by the broker and the extension of its obligation to counsel.The legal analysis of the consumption process of financial products supposes the combination of all these elements. On the one hand, before and during the formation of links, governments attempt to restore balance in relations between professionals and consumers. The latter being deemed economically weak, do not have the same information on financial products and services are sought by various means and technologies they have no control, and must make choices between various products, and more complex. On the other hand, during the execution of services, in addition to commodity risk, or non-performance of contractual obligations, other risks may arise from failures in market functioning and behavior of its actors. Therefore a means of strengthening consumer protection is needed to overcome the shortcomings of current means that they have to obtain compensation and seek compensation.